1. Using the R Tidyverse packages
Simon Andrews v

2. What's wrong with R? Stupid legacy design choices Inconsistent API
Difficult to follow code logical flow
Multiple ways to model data

3. Stupid Design Choices data.frame( gene=c("ABC10","DEF10","GHI10"),
"gene name"=c("ABC10 gene","DEF10 gene","GHI10 Gene"),
value=c(10,12,14),
row.names=1
) -> some.data

4. Stupid Design Choices class(some.data$name) [1] "factor"
some.data["ABC1",]
name value
ABC10 ABC10 gene 10

5. Stupid Design Choices small.data
interesting.samples <- c("sample1","sample2")
small.data[,interesting.samples]
sample1 sample2
ABC
DEF
GHI
interesting.samples <- c("sample3")
[1]

6. Inconsistent API barplot(1:3, horiz=TRUE)
stripchart(1:3,vertical = FALSE)
min(DATA)
log(DATA,[other options])
substr(DATA,[other options])
grep(pattern,DATA,[other options])

7. Difficult to follow logical flow6 5 4 2 1 3
sqrt(min(log(abs(some.data)+1,base=2)))
Which function does
this belong to?

8. Tidyverse Collection of R packages
Collection of R packages
Aims to fix many of core R's structural problems
Common design and data philosophy
Designed to work together, but integrate seamlessly with other parts of R

9. Tidyverse Packages Tibble - data storage
ReadR - reading data from files
TidyR - Model data correctly
DplyR - Manipulate and filter data
Ggplot2 - Draw figures and graphs

10. Tidyverse Philosophies
All data stored in a tibble (like a data frame)
Make all functions perform a single, simple operation
Function names should be descriptive - normally a verb
All functions take a tibble as their first argument, allowing 'pipes' to be constructed
All (well most) functions return a tibble
Use functional programming (not OO)
Don't modify existing data - create a modified copy

11. Using Tidyverse

12. Things to cover today Installing Tidyverse
Reading files and using tibbles
Restructuring data into ‘tidy’ format
Transforming tidy data
Subsetting and filtering
Grouping and Summarising
Extending and Joining
Final exercise

13. Installation and calling
install.packages("tidyverse")
library("tidyverse")
-- Attaching packages tidyverse
v ggplot v purrr
v tibble v dplyr
v tidyr v stringr 1.3.1
v readr v forcats 0.3.0

14. Reading Files with readr
Provides functions which mirror R's read.table
read_csv("file.csv")
read_tsv("file.tsv")
read_delim("file.txt")
read_fwf("file.txt",col_positions=c(1,3,6))

15. Reading files with readr
Output is always a tibble
No name translation
Guesses appropriate formats
Says what formats were guessed (can get it wrong though!)
No string to factor conversion

16. Reading files with readr
> read_tsv("trumpton.txt") -> trumpton
Parsed with column specification:
cols(
LastName = col_character(),
FirstName = col_character(),
Age = col_double(),
Weight = col_double(),
Height = col_double() )
> trumpton
# A tibble: 7 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Hugh Chris
2 Pew Adam
3 Barney Daniel
4 McGrew Chris
5 Cuthbert Carl
6 Dibble Liam
7 Grub Doug

17. Fixing guessed columns
> read_csv("Child_Variants.csv") -> child_tbl
Parsed with column specification:
cols(
CHR = col_double(),
POS = col_double(),
dbSNP = col_character(),
REF = col_character(),
ALT = col_character(),
QUAL = col_double(),
GENE = col_character(),
ENST = col_character(),
MutantReads = col_double(),
COVERAGE = col_double(),
MutantReadPercent = col_double() )
Warning: 477 parsing failures.
row col expected actual file
25346 CHR a double MT 'Child_Variants.csv'
25347 CHR a double MT 'Child_Variants.csv'
25348 CHR a double MT 'Child_Variants.csv'
25349 CHR a double MT 'Child_Variants.csv'
25350 CHR a double MT 'Child_Variants.csv'
See problems(...) for more details.
Types are guessed on first 1000 lines
Warnings for later mismatches
Values converted to NA

18. Fixing guessed columns
> read_csv(
"Child_Variants.csv",
col_types=cols(CHR=col_character())
) -> child_tbl
# A tibble: 25,822 x 11
CHR POS dbSNP REF ALT QUAL GENE ENST MutantReads COVERAGE
<chr> <dbl> <chr> <chr> <chr> <dbl> <chr> <chr> <dbl> <dbl>
A G OR4F5 ENST~
rs75~ A G OR4F5 ENST~
A T OR4F5 ENST~
rs75~ T C OR4F5 ENST~
rs66~ T C SAMD~ ENST~
rs22~ G A NOC2L ENST~
rs38~ A G NOC2L ENST~
rs37~ T C NOC2L ENST~
rs37~ T C NOC2L ENST~
rs13~ G C NOC2L ENST~
# ... with 25,812 more rows, and 1 more variable: MutantReadPercent <dbl>

19. Tibbles vs Data Frames Most operations are compatible my.tibble$column
Any function taking data frame as argument

20. Tibbles vs Data Frames Tibbles print more nicely
Tell you the column types
Tell you the overall dimensions
Omit columns which don't fit
Truncate values which don’t fit
Show fewer rows by default (10 vs 1000)
Tibbles don't support row names
Causes issues with some packages, especially BioConductor

21. Printing tibbles Tells you the column types
# A tibble: 25,822 x 11
CHR POS dbSNP REF ALT QUAL GENE ENST MutantReads COVERAGE
<chr> <dbl> <chr> <chr> <chr> <dbl> <chr> <chr> <dbl> <dbl>
A G OR4F5 ENST~
rs75~ A G OR4F5 ENST~
A T OR4F5 ENST~
rs75~ T C OR4F5 ENST~
rs66~ T C SAMD~ ENST~
rs22~ G A NOC2L ENST~
rs38~ A G NOC2L ENST~
rs37~ T C NOC2L ENST~
rs37~ T C NOC2L ENST~
rs13~ G C NOC2L ENST~
# ... with 25,812 more rows, and 1 more variable: MutantReadPercent <dbl>
Tells you the column types
Tells you the overall dimensions
Omits columns which don't fit
Truncates values which don’t fit
Shows fewer (10 vs 1000) rows by default

22. Creating tibbles Any read method from readr
Using as_tibble() on a data frame
Using tibble() from vectors

23. Exercise 1 Reading Data into Tibbles

24. "Tidy Data" Many ways to store the same data
# A tibble: 3 x 8
Gene Chr Start End Sample1_WT Sample2_WT Sample3_KO Sample4_KO
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Gnai
2 Pbsn
3 Cdc
Many ways to store the same data
Different structures need different code to process them
Good to have a common design structure

25. Wide vs Long Wide format Rows are measures Samples are columns
Implied (often incorrectly)linkage between measures on the same row for different samples
Works OK when measures are paired and equal
Tricky to alter as you need to select columns first

26. Wide vs Long Long format is the standard "Tidy" format
Every row is a single observation
Every column should be a different type of annotation or measure
Should never have the same measure for difference samples in multiple columns
Slightly more complex for paired studies
More flexible and consistent generally.

27. Converting to "Tidy" format
# A tibble: 3 x 8
Gene Chr Start End Sample1_WT Sample2_WT Sample3_KO Sample4_KO
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Gnai
2 Pbsn
3 Cdc
Put all measures into a single column
Add a 'sample' and 'genotype' column
Duplicate the gene information as required
Or separate it into a different table

28. Converting to "Tidy" format
# A tibble: 3 x 8
Gene Chr Start End Sample1_WT Sample2_WT Sample3_KO Sample4_KO
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Gnai
2 Pbsn
3 Cdc
non.normalised %>%
gather(sample, value, -Gene,-Chr,-Start,-End) %>%
separate(sample,into=c("sample","genotype"),sep="_")

29. Converting to "Tidy" format
# A tibble: 12 x 7
Gene Chr Start End sample genotype value
<chr> <dbl> <dbl> <dbl> <chr> <chr> <dbl>
1 Gnai Sample1 WT
2 Pbsn Sample1 WT
3 Cdc Sample1 WT
4 Gnai Sample2 WT
5 Pbsn Sample2 WT
6 Cdc Sample2 WT
7 Gnai Sample3 KO
8 Pbsn Sample3 KO
9 Cdc Sample3 KO
10 Gnai Sample4 KO
11 Pbsn Sample4 KO
12 Cdc Sample4 KO

30. Converting to "Tidy" format
Can clean up duplicated information
# A tibble: 12 x 4
Gene sample genotype value
<chr> <chr> <chr> <dbl>
1 Gnai3 Sample1 WT
2 Pbsn Sample1 WT
3 Cdc45 Sample1 WT
4 Gnai3 Sample2 WT
5 Pbsn Sample2 WT
6 Cdc45 Sample2 WT
7 Gnai3 Sample3 KO
8 Pbsn Sample3 KO
9 Cdc45 Sample3 KO
10 Gnai3 Sample4 KO
11 Pbsn Sample4 KO
12 Cdc45 Sample4 KO
# A tibble: 3 x 4
Gene Chr Start End
<chr> <dbl> <dbl> <dbl>
1 Gnai
2 Pbsn
3 Cdc
These can be recombined later on as needed.

31. Tidying operations gather spread separate unite
Takes multiple columns of the same type and puts them into a pair of key-value columns
spread
Takes a key-value column pair and spreads them out to multiple columns of the same type
separate
Splits a delimited column into multiple columns
unite
Combines multiple columns into one

32. Using gather gather(gather.data,sample,value,-genes)
gather(data,key_name,value_name, columns)
gather(gather.data,sample,value,-genes)
# A tibble: 12 x 3
genes sample value
<chr> <chr> <dbl>
1 Nanog sample
2 Sfi1 sample
3 GAPDH sample
4 Nanog sample
5 Sfi1 sample
6 GAPDH sample
7 Nanog sample
8 Sfi1 sample
9 GAPDH sample
10 Nanog sample
11 Sfi1 sample
12 GAPDH sample
> gather.data
# A tibble: 3 x 5
genes sample1 sample2 sample3 sample4
<chr> <dbl> <dbl> <dbl> <dbl>
1 Nanog
2 Sfi
3 GAPDH
Column selections can be positive (sample1,sample2,sample3)
or negative (-genes)

33. Using spread spread(spread.data,dimension,measures)
spread(data,key_col,value_col)
spread(spread.data,dimension,measures)
samples depth height width
<chr> <int> <int> <int>
1 sample
2 sample
> spread.data
# A tibble: 6 x 3
samples dimension measures
<chr> <chr> <int>
1 sample1 height
2 sample1 width
3 sample1 depth
4 sample2 height
5 sample2 width
6 sample2 depth

34. Using separate separate(data,col,into_vector,sep)
separate(sep.data,size,into=c("h","w"),sep="x",convert = TRUE)
# A tibble: 4 x 3
standard h w
<chr> <int> <int>
1 SVGA
2 XGA
3 WXGA
4 FHD
> sep.data
# A tibble: 4 x 2
standard size
<chr> <chr>
1 SVGA x600
2 XGA x768
3 WXGA x720
4 FHD x1080
convert=TRUE means that the type for the new columns will be re-guessed since it might well change.

35. Using unite unite(data,cols,colname,sep)
unite(unite.data,chr,start,strand,col="locus",sep=":")
# A tibble: 3 x 2
locus value
<chr> <dbl>
1 1:12876:
2 2: :
3 X: :- 72.1
> unite.data
# A tibble: 3 x 4
chr start strand value
<chr> <dbl> <chr> <dbl>
3 X

36. Multiple operations gather the sample columns into sample and value
# A tibble: 3 x 8
Gene Chr Start End Sample1_WT Sample2_WT Sample3_KO Sample4_KO
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Gnai
2 Pbsn
3 Cdc
gather the sample columns into sample and value
Split the sample column into sample and genotype with separate

37. Multiple operations
gather(non.norm,sample,value,-Gene,-Chr,-Start,-End) -> n1
separate(n1, sample, into=c("sample","geno"),sep="_") -> n2
separate(
gather(
non.normalised,
sample,
value,
-Gene,-Chr,-Start,-End ),
into=c("sample","genotype"),
sep="_"
) -> norm2

38. The %>% operator (pipe)
All* tidyverse functions take a tibble as the first argument and return a modified tibble
Often need to 'chain' several functions together.
The %>% operator takes the argument on its left and makes it the first argument to the function on its right
Correct logical flow to the code
Easier to read and interpret
Cleaner (fewer intermediate variables, less typing)
*Well nearly all, anyway.

39. The %>% operator (pipe)
data %>% head()
head(data)
tibble(a=1:10,b=21:30) %>% nrow()
*Well nearly all, anyway.

40. Multiple operations non.norm %>%
# A tibble: 3 x 8
Gene Chr Start End Sample1_WT Sample2_WT Sample3_KO Sample4_KO
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Gnai
2 Pbsn
3 Cdc
non.norm %>%
gather(sample,value,-Gene,-Chr,-Start,-End) %>%
separate(sample,into=c("sample","geno"),sep="_")

41. Exercise 2 Restructuring data into ‘tidy’ format

42. Transforming data with dplyr

43. Functions in dplyr
Replacement for core selecting, sorting and filtering
Operations can be split into groups
Subsetting or filtering a tibble
Grouping and summarising variables
Extending tibbles by adding data

44. Subsetting and Filtering
slice select rows by position
select select columns by name/position
arrange sort rows
distinct deduplication
filter functional selection of rows

45. Trumpton # A tibble: 7 x 5 LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Hugh Chris
2 Pew Adam
3 Barney Daniel
4 McGrew Chris
5 Cuthbert Carl
6 Dibble Liam
7 Grub Doug

46. Using slice # A tibble: 3 x 5 LastName FirstName Age Weight Height
slice(data,rows)
trumpton %>% slice(1,4,7)
# A tibble: 3 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Hugh Chris
2 McGrew Chris
3 Grub Doug
Note, rows can be separate (as above) or combined into one vector

47. Using select # A tibble: 7 x 3 LastName Age Height
select(data,cols)
trumpton %>% select(LastName,Age,Height)
# A tibble: 7 x 3
LastName Age Height
<chr> <dbl> <dbl>
1 Hugh
2 Pew
3 Barney
4 McGrew
5 Cuthbert
6 Dibble
7 Grub

48. Using select # A tibble: 7 x 4 FirstName Age Weight Height
select(data, -unwanted_cols)
trumpton %>% select(-LastName)
# A tibble: 7 x 4
FirstName Age Weight Height
<chr> <dbl> <dbl> <dbl>
1 Chris
2 Adam
3 Daniel
4 Chris
5 Carl
6 Liam
7 Doug

49. Defining Selected Columns
Common rules used throughout tidyverse.
Single definitions (name, position or function)
Positive weight, height, length, 1, 2, 3, last_col()
Negative -chromosome, -start, -end, -1, -2, -3
Range selections
3:5 -(3:5) height:length -(height:length)
Functional selections (positive or negative)
starts_with() -starts_with()
ends_with() -ends_with()
contains() -contains()
matches() -matches

50. Using select helpers > child.variants
CHR POS dbSNP REF ALT QUAL GENE ENST MutantReads COVERAGE MutantReadPercent
> child.variants %>% select(REF,COVERAGE)
REF COVERAGE
> child.variants %>% select(-CHR, -ENST)
POS dbSNP REF ALT QUAL GENE MutantReads COVERAGE MutantReadPercent
> child.variants %>% select(5:last_col())
ALT QUAL GENE ENST MutantReads COVERAGE MutantReadPercent
> child.variants %>% select(POS:GENE)
POS dbSNP REF ALT QUAL GENE
> child.variants %>% select(-(POS:GENE))
CHR ENST MutantReads COVERAGE MutantReadPercent
> child.variants %>% select(starts_with("Mut"))
MutantReads MutantReadPercent
> child.variants %>% select(-ends_with("t",ignore.case = TRUE))
CHR POS dbSNP REF QUAL GENE MutantReads COVERAGE
> child.variants %>% select(contains("Read"))
MutantReads MutantReadPercent

51. Using arrange # A tibble: 7 x 5 LastName FirstName Age Weight Height
arrange(data,cols)
trumpton %>% arrange(Age)
# A tibble: 7 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Barney Daniel
2 Hugh Chris
3 Cuthbert Carl
4 Grub Doug
5 Pew Adam
6 Dibble Liam
7 McGrew Chris

52. Using arrange # A tibble: 7 x 5 LastName FirstName Age Weight Height
arrange(data,cols)
trumpton %>% arrange(desc(Age))
# A tibble: 7 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 McGrew Chris
2 Dibble Liam
3 Pew Adam
4 Grub Doug
5 Cuthbert Carl
6 Hugh Chris
7 Barney Daniel

53. Using distinct # A tibble: 6 x 5 LastName FirstName Age Weight Height
distinct(data,cols, .keep_all=TRUE)
trumpton %>% distinct(FirstName,.keep_all=TRUE)
# A tibble: 6 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Hugh Chris
2 Pew Adam
3 Barney Daniel
4 Cuthbert Carl
5 Dibble Liam
6 Grub Doug
Note, “keep_all” has a dot before it

54. Using filter # A tibble: 6 x 5 LastName FirstName Age Weight Height
filter(data,conditions)
trumpton %>% filter(Height>170 & Age < 30)
# A tibble: 6 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Hugh Chris
2 Cuthbert Carl

55. Clashes with other packages
Many other packages define a function called select (eg MASS)
In R, the last package loaded, wins. Tidyverse usually loses since it’s loaded at the start!
In these cases you need to specify the package as well as the function:
dplyr::select()
Can affect other functions too

56. Clashes with other packages
> library(MASS)
Attaching package: ‘MASS’
The following object is
masked from ‘package:dplyr’:
select
> select
function (obj)
UseMethod("select")
<bytecode: 0x >
<environment: namespace:MASS>
> dplyr::select
function (.data, ...) { }
<bytecode: 0x b90>
<environment: namespace:dplyr>

57. Exercise 3 Filtering and selecting with dplyr

58. Grouping and Summarising
group_by sets groups for summarisation
ungroup removes grouping information
summarise collapse grouped variables
count count grouped variables

59. Grouping and Summarising Workflow
Load a tibble with repeated values in one or more columns
Use group_by to select all of the categorical columns you want to combine to define your groups
Run summarise saying how you want to combine the quantitative values
Run ungroup to remove any remaining group information

60. Grouping and Summarising Workflow
Load a tibble with repeated values in one or more columns
Use group_by to select all of the categorical columns you want to combine to define your groups
Run summarise saying how you want to combine the quantitative values
Run ungroup to remove any remaining group information

61. Grouping and Summarising
> group.data
# A tibble: 8 x 5
Sample Genotype Sex Height Length
<dbl> <chr> <chr> <dbl> <dbl>
WT M
WT F
WT F
WT M
KO M
KO F
KO F
KO M
Want to get average Height and Length for each sex, but still split by genotype

62. Grouping and Summarising Workflow
Load a tibble with repeated values in one or more columns
Use group_by to select all of the categorical columns you want to combine to define your groups
Run summarise saying how you want to combine the quantitative values
Run ungroup to remove any remaining group information

63. Grouping and Summarising
Discard
Group
Mean
Median
Sample Genotype Sex Height Length
<dbl> <chr> <chr> <dbl> <dbl>
Categorical
Quantitative
Want to get average Height and Length for each combination of sex and genotype

64. Grouping and Summarising
group.data %>% group_by(Genotype,Sex)
# A tibble: 8 x 5
# Groups: Genotype, Sex [4]
Sample Genotype Sex Height Length
<dbl> <chr> <chr> <dbl> <dbl>
WT M
WT F
WT F
WT M
KO M
KO F
KO F
KO M

65. Grouping and Summarising Workflow
Load a tibble with repeated values in one or more columns
Use group_by to select all of the categorical columns you want to combine to define your groups
Run summarise saying how you want to combine the quantitative values
Run ungroup to remove any remaining group information

66. Grouping and Summarising
group.data %>%
group_by(Genotype,Sex) %>%
summarise(Height2=mean(Height),Length=median(Length))
# A tibble: 4 x 4
# Groups: Genotype [2]
Genotype Sex Height2 Length
<chr> <chr> <dbl> <dbl>
1 KO F
2 KO M
3 WT F
4 WT M

67. Grouping and Summarising Workflow
Load a tibble with repeated values in one or more columns
Use group_by to select all of the categorical columns you want to combine to define your groups
Run summarise saying how you want to combine the quantitative values
Run ungroup to remove any remaining group information

68. Ungrouping
A summarise operation removes the the last level of grouping (“sex” in our worked example)
Other levels of grouping (“Genotype”) remain annotated on the data, so you could do an additional summarisation if needed
If you’re not going to use them it’s a good idea to use ungroup to remove remaining groups so they don’t interfere with other operations

69. Exercise 4 Grouping and Summarising

70. Extending tibbles mutate compute a new column
add_row adds an additional row
bind_rows join two tibbles by row
add_column adds an additional column
bind_cols join two tibbles by column
rename renames an existing column

71. Using mutate mutate(data,newcol=oldcol1+oldcol2)
trumpton %>% mutate(bmi=Weight/(Height/100)^2)
# A tibble: 7 x 6
LastName FirstName Age Weight Height bmi
<chr> <chr> <dbl> <dbl> <dbl> <dbl>
1 Hugh Chris
2 Pew Adam
3 Barney Daniel
4 McGrew Chris
5 Cuthbert Carl
6 Dibble Liam
7 Grub Doug

72. Using add_row add_row(data,key1=value1, key2=value2) trumpton %>%
LastName="Andrews",FirstName="Simon",
Age=39,Weight=80,Height=190 )
# A tibble: 8 x 5
LastName FirstName Age Weight Height
<chr> <chr> <dbl> <dbl> <dbl>
1 Hugh Chris
2 Pew Adam
3 Barney Daniel
4 McGrew Chris
5 Cuthbert Carl
6 Dibble Liam
7 Grub Doug
8 Andrews Simon

73. Using bind_rows # A tibble: 6 x 2 gene value <chr> <dbl>
bind_rows(data1,data2)
bind_rows(some.data,some.more.data)
> some.data
# A tibble: 3 x 2
gene value
<chr> <dbl>
1 ABC
2 DEF
3 GHI
> some.more.data
1 JKL
2 MNO
3 PQR
# A tibble: 6 x 2
gene value
<chr> <dbl>
1 ABC
2 DEF
3 GHI
4 JKL
5 MNO
6 PQR

74. Using add_column # A tibble: 3 x 3 gene value value2
add_column(data,newcol=new_vector)
add_column(some.data,value2=c(16,18,20))
# A tibble: 3 x 3
gene value value2
<chr> <dbl> <dbl>
1 ABC
2 DEF
3 GHI

75. Using bind_cols # A tibble: 3 x 4 gene value gene1 value1
bind_cols(tibble1,tibble2)
bind_cols(some.data,some.more.data)
> some.data
# A tibble: 3 x 2
gene value
<chr> <dbl>
1 ABC
2 DEF
3 GHI
> some.more.data
1 JKL
2 MNO
3 PQR
# A tibble: 3 x 4
gene value gene1 value1
<chr> <dbl> <chr> <dbl>
1 ABC JKL
2 DEF MNO
3 GHI PQR

76. Using rename # A tibble: 3 x 2 gene_name value <chr> <dbl>
rename(data,new_name=name)
rename(some.data, gene_name = gene)
> some.data
# A tibble: 3 x 2
gene value
<chr> <dbl>
1 ABC
2 DEF
3 GHI
# A tibble: 3 x 2
gene_name value
<chr> <dbl>
1 ABC
2 DEF
3 GHI

77. Joining tibbles x and y left_join join matching values from y into x
right_join join matching values of x into y
inner_join join x and y keeping only rows in both
full_join join x and y keeping all values in both

78. Join types left_join right_join inner_join full_join > join1
name count percentage
1 Simon
2 Steven NA
3 Felix
> join1
name count
1 Simon
2 Steven 6
3 Felix
> join2
name percentage
1 Felix
2 Anne
3 Simon
right_join
name count percentage
1 Felix
2 Anne NA
3 Simon
inner_join
name count percentage
1 Simon
2 Felix
full_join
name count percentage
1 Simon
2 Steven NA
3 Felix
4 Anne NA

79. Joining options by specify the columns to join on
Simple name if it’s the same between both
by=“gene”
Paired names if they differ between x and y
by=c(“gene” = “gene_name”)
suffix the text suffix for duplicated column names

80. Exercise 5 Extending and Joining